Upload
others
View
2
Download
0
Embed Size (px)
Citation preview
MAND TRAINING AND PROMPT FADING: TEACHING YES-NO MANDING
By
Jessie Kelly
A Project Presented to
The Faculty of Humboldt State University
In Partial Fulfillment of the Requirements for the Degree
Master of Arts in Psychology: Academic Research
Dr. Christopher Walmsley, Committee Chair
Ian Jewett, Committee Member
Dr. Chris Aberson, Program Graduate Coordinator
May 2020
ii
Abstract
MAND TRAINING AND PROMPT FADING: TEACHING YES-NO MANDING
Jessie Kelly
Language is a collection of verbal skills which allows the speaker to have some control
over how others interact with them. Saying the words “yes” and “no” are fundamental
language abilities; The words impact the speaker’s environment by communicating
acceptance or rejection of stimuli and experiences. Caregivers, teachers, and friends may
have to make choices for an individual who cannot respond to yes-no questions. This
skill deficit deprives the individual of autonomy and can lead to challenging behaviors
occurring in place of clear communication. Applied Behavior Analysis offers techniques
to teach language skills such as saying “yes” and “no” in response to verbal and
environmental stimuli. This study used the techniques of mand training in discrete trial
format while systematically fading prompts to teach two individuals with autism to
correctly respond to the question “do you want this?” paired with the presentation of a
highly preferred or lesser preferred item. Correct responding was predetermined with
paired stimulus preference assessments by identifying “yes” items and “no” items. The
results from this study suggest that these procedures are effective at producing correct
responding to trained and untrained stimuli, thereby allowing the individual to
functionally communicate their wants and needs.
iii
Acknowledgements
Thank you to Ian Jewett, Liann Carrillo, Erin Buchan, and Kier Flicker at
Multiplicity Therapeutic Services for approving and assisting with this project from
beginning to end.
Thank you to Professor Walmsley at Humboldt State University for supporting
this project and prompting consistent collaboration and meetings, despite skunk attacks
and pandemics.
iv
Table of Contents
Abstract ............................................................................................................................... ii
Acknowledgements ............................................................................................................ iii
List of Figures ..................................................................................................................... v
Introduction ......................................................................................................................... 1
Methodology ....................................................................................................................... 6
Participants ...................................................................................................................... 6
Inclusionary criteria .................................................................................................... 7
Materials ......................................................................................................................... 7
Measures and Interobserver Agreement ......................................................................... 8
Research Procedures ....................................................................................................... 9
Setting ......................................................................................................................... 9
Caregiver interview ..................................................................................................... 9
Preference assessments ............................................................................................. 10
Baseline phase ........................................................................................................... 11
Intervention ............................................................................................................... 11
Research Design ........................................................................................................... 14
Results ............................................................................................................................... 15
Discussion ......................................................................................................................... 17
References ......................................................................................................................... 22
v
List of Figures
Figure 1. Jacob and Maya’s results across baseline, prompt fading, and generalization
phases. ............................................................................................................................... 16
1
Introduction
Autism Spectrum Disorder (ASD) is a neurodevelopmental disorder characterized
by deficits in social communication and restricted, repetitive patterns of behavior. The
diagnosis is separated into severity levels according to the amount of one-on-one support
the individual may require (American Psychiatric Association, 2013). For example,
moderate to intense support may be needed to assist an individual with daily functioning
and effective communication. The individual may lack the ability to gesture or point,
respond to their name or simple directions, and may repeat words or phrases exactly as
heard but out of context. These considerable skill deficits can hinder appropriate
expression of needs, rejection of unwanted experiences, and initiation or responding to
others (Buron, Wolfberg, & Grey, 2014).
In the mid-1900s, investigators began to experiment with various behavior
analytic treatments to address the language acquisition challenges that individuals with
autism face. These treatments are now trusted, evidence-based interventions for children,
adolescents, and adults because they rely on established behavior principles and rigorous
experimental research (Smith, 1999). Applied Behavior Analysis (ABA) is a type of
behavioral science which focuses on the discovery and manipulation of environmental
variables that reliably influence behaviors. Interventions based on ABA target interfering
behaviors for reduction and socially important skills for acquisition while considering the
ethical implications of each treatments (Cooper, Heron, & Heward, 2007).
2
Verbal behavior acquisition is a primary focus of ABA treatments. Fundamental
components of language were identified in B.F. Skinner’s book, Verbal Behavior (1957),
and are referred to as the elementary verbal operants (tact, mand, intraverbal, echoic,
textual, and transcription). An operant is a class of behavior which is influenced by
consequence events in the environment (i.e. a learned behavior). A mand is a verbal
operant to which a particular response is reinforced by a specific consequence and is
therefore under the control of an establishing operation relevant to that consequence. An
establishing operation is a behavioral effect that momentarily alters the reinforcing
effectiveness of other events and alters the frequency of the kind of responses that have
been reinforced by those other events (Michael, 1988). For example, an individual may
ask for a snack because they are hungry. It can be said that that are experiencing food
deprivation (establishing operation) and therefore the snack’s reinforcing effects are
altered (increased) and this evokes various responses associated with gaining access to a
snack (manding “I want a snack please”). Plavnick and Ferreri (2012) describe a mand as
allowing the speaker to have some control over the way other people in their environment
provide reinforcement and for this reason it is suggested to be the initial training target
for verbal behavior. Jennett, Harris, and Delmolino (2008) further state that when the
speaker can make their needs and wants known by manding, behaviors that have
previously interfered with the production of functional language will decrease and verbal
behavior will increase. The vision of helping individuals with autism meet their needs
with a recognizable and generalizable form (verbal behavior) has led to the development
of specific teaching protocols based in ABA technology.
3
Mand training is the application of reinforcement specific to the mand. Mand
training protocols may take advantage of naturally occurring stimuli and the resulting
reinforcement contingencies (natural environment training: NET) or may contrive the
environment to increase the chances of successful responding (discrete trial training:
DTT). DTT is a controlled type of teaching in which an opportunity to mand (respond) is
created in the environment (Cooper, et al., 2007). Although a combination of DTT and
NET achieves a more complete language repertoire, previous research supports the
hypothesis that specific manding skills are acquired to individualized criteria when taught
using DTT alone (Sundberg & Michael, 2001). This study focused on the latter method of
training in an effort to provide further empirical support for mand acquisition using DTT.
The response target for mand training will depend on the learner’s verbal
repertoire and the type of mand that will provide them with the most opportunities for
reinforcement in their environment. In view of this information, this study focused on
“yes” and “no” manding as the target responses to teach learners. Emitting “yes” and
“no” in the presence of desired or undesired environmental stimuli (e.g. food, toys, a
tangible item) is considered a basic language skill because of the potential for those
responses to have multiple effects on the environment (Shillingsburg, Kelley, Roane,
Kisamore, & Brown, 2009). For example, a child responding “no” to environmental
stimuli may indicate to remove an item, allow a break from a task, or to give the child
space when they want to be alone. Responding “yes” to environmental stimuli can
indicate the acceptance of an item or the desire for continued access to an activity. Once
yes-no manding skills are acquired in a controlled environment, responding may
4
generalize to naturally occurring stimuli because of the abundant opportunities for
positive and negative reinforcement as a consequence of manding “yes” and “no.”
The study used DTT to teach yes-no manding to individuals with autism based on
the results of a preference assessment. Research by Shillingsburg, et al. (2009) used
preference assessments to establish correct yes-no responses to stimuli across verbal
operants (tacts, mands, and intraverbals) and DTT format for teaching and generalization
trials. Earlier studies such as Neef, Walters, and Egel (1984) and Hung (1980) used a
combination of DTT and NET to teach yes-no manding but did not use preference
assessments to identify stimuli as “yes” and “no” items which left correct and incorrect
scoring up to subjective judgment. Many individuals with autism are unable to self-report
their preferences. Without clear communication from the individual, others in the
environment may guess what they want or choose for them out of necessity (Mangum,
Fredrick, Pabico, & Roane, 2012). This study used a paired stimulus preference
assessment to pre-identify highly preferred (“yes”) items in contrast to lesser preferred
(“no”) items. Pre-identification allowed for objective scoring of correct and incorrect
responses which indicated if the learner had acquired the skills of yes-no manding. This
study attempted to fill the gaps in earlier research and extend recent research by adding
empirical support for the use of a preference assessment to pre-determine highly
preferred versus lesser preferred stimuli.
Some studies failed to test for generalization of yes-no mand responses with
untrained stimuli. These studies showed that yes-no responses can be trained for
preferred and lesser preferred food items, but trained mand responses may not result with
5
untrained stimuli (Hung, 1980; Neef, et al., 1984). This study used training techniques
(errorless learning and prompt fading with a visual support) that have not been previously
implemented in published research. This study found that the use of multiple training
techniques resulted in the participants’ ability to generalize yes-no responses to untrained
stimuli.
6
Methodology
Participants
The participants in this study were recruited from an ABA agency in Humboldt
County, California. Before recruitment took place, a research proposal was submitted to
the Humboldt State University Institutional Review Board and was approved on May
15th, 2019 with the approval number IRD 18-191. Next, the primary researcher and the
agency’s clinical director identified potential participants through clinical reports and
observations which indicated that they struggled with responding to yes-no questions.
Three individuals were initially selected but the participation of one was discontinued due
to revocation of assent during teaching trials. The remaining participants were Jacob, a
three-year-old boy and Maya, a twenty-year-old woman, both diagnosed with ASD.
The clinical director began recruitment by meeting with the caregivers, describing
the study, then asking if they would like to participate. During this meeting, each family
was informed that participation was voluntary, and their choice would not affect current
or future therapy services provided by the agency. If the caregivers agreed to participate,
the clinical director reviewed the informed consent document with them and acquired a
signature by at least one legal guardian.
Clinical staff from the agency were recruited to conduct reliability checks
throughout this study. The researcher informed each clinician of the project objectives,
the procedures they would partake in, and explained that there were no risks to their
7
person or employment whether they chose to participate or not. They verbally agreed to
participate, then signed an informed consent document.
Inclusionary criteria. The researcher interacted with the individuals and assessed
if they could echo “yes” and “no” when the words were said in their immediate
proximity. The researcher also confirmed with caregivers that the individuals struggled
with responding to questions with a vocal “yes” or “no” response. If the individual could
repeat “yes” and “no” and they were reported to incorrectly respond to yes-no questions,
they were included for participation in this research.
Materials
Food was used during the preferences assessments, baseline sessions, teaching
trials, and generalization probes. The edible items were provided by the researcher and
brought to each session in plastic food containers. Containers were placed behind a large
cardboard barrier to prevent distraction while testing specific items. Participants were
provided with a paper napkin, a plastic bowl, and water or a preferred drink. During the
teaching trials, a paper with the words “yes” and “no” printed in black ink against a green
or red background was used to prompt correct responding when the procedures required
its use.
The researcher recorded responses on pre-made paper data sheets. These papers
were kept with the researcher in a binder until they were transported to the university
research lab and placed in a secure file cabinet with the informed consent documents.
8
Electronic forms were stored by the researcher on a flash drive which was also kept in the
university’s secure file cabinet.
Measures and Interobserver Agreement
The primary dependent measure in this study was correct and incorrect yes and no
responses to the presented food item and the verbal stimulus “do you want this?” The
second dependent measure was the ranked ordering of food items from the preference
assessments. The third dependent measure was responses to the social validity survey.
During baseline, teaching trials, and the generalization probes, correct responses
were recorded if the participant responded to the presented item and verbal stimulus with
the appropriate response within 5-sec. Incorrect responses were recorded if the participant
did not respond within 5-sec or did not respond appropriately. The data was summarized
as the percentage of correct responding per session.
Preference assessment responses were recorded as the item chosen per trial. Each
assessment consisted of 45 trials which assessed 10 food items. After the trails
concluded, the researcher counted the number of times that each item was chosen and
divided that number by the total amount of trials. The final score per item was the
percentage of time it was chosen, and items were ranked from most to least chosen to
indicate highly preferred to lesser preferred.
At the end of this study, the researcher asked both families to fill out a survey to
assess their overall satisfaction. There were three 0-5 ranking scales to which a score of
zero indicated low satisfaction while a score of five indicated high satisfaction. Both
9
families scored five for each ranked item. There were two additional questions, which
both families scored “yes,” indicating their agreement that they benefited from
participating in this study and the teaching procedures were appropriate for each
participant’s learning style.
A clinical staff member independently collected inter-observer agreement (IOA)
data during a mean of 46.5% of sessions. Agreement was defined as both the researcher
and the additional observer agreeing on the occurrence or nonoccurrence of a correct or
incorrect response. For both participants, point-by-point agreement scores were 100% per
session and across sessions, which lead to a final IOA score of 100% agreement
throughout this study.
Research Procedures
Setting. For Jacob, the procedures took place in his home. The researcher sat on
the living room floor with Jacob during all trials while his mother was in a nearby room.
The materials were near the researcher and behind a cardboard partition. For Maya, the
caregiver interview took place in her home, while the other procedures took place at the
agency’s office. The office room contained one desk, a couch, and a small table. The
materials were placed on the desk, behind the partition, and the researcher sat across from
Maya, who sat on the couch.
Caregiver interview. The researcher met with the families and asked about
known preferred and non-preferred foods and items. Based on the information, food was
10
determined to be highly preferred over toys or activities for both participants and was
therefore selected for the preference assessments.
Preference assessments. Structured preference assessments are used to detect
highly preferred and lesser preferred stimuli. Stimuli that are interacted with more often
are deemed preferred and stimuli that are interreacted with less often are deemed lesser
preferred (Kang et al., 2013). Food items were used to establish highly preferred and
lesser preferred stimuli. Data from the preference assessments informed the correct
responses during baseline, teaching trials, and the generalization probes.
The type of assessment used in this study was a paired stimulus preference
assessment. First, the participant sampled each food item for approximately 30-sec, or
until it was consumed. During the trials, the participant was presented with two food
items at once and told to “pick one.” The item chosen was recorded on a data sheet and
the participant had approximately 10 to 30-sec to interact with the food. This process was
repeated until all 10 pre-selected items were tested against each other over a total of 45
trials. If the participant did not make a choice when told to “pick one,” the items were
withdrawn after approximately 5-sec, then represented with the instruction of “pick one”
once more. If the participant did not make a choice after the second presentation of the
same items, the trial ended and was crossed out on the data sheet, then the next two items
were presented. When this occurred, the final scores were adjusted based on the amount
of trials that an item was selected. The same procedure was followed if the participant
reached for both items at once or said “all done” when presented with a hypothesized
lesser-preferred item. If the participant selected an item but did not eat it, or chewed it but
11
spit it out, the item selected was still indicated on the data sheet but a note was made next
to it. This did not factor into the final assessment scores.
The final scores were written in descending order of most chosen to least chosen
item. The top five items were chosen more often than the bottom five and therefore
became the “yes” items while the lower ones became the “no” items for baseline and
teaching trials. A second preference assessment, which followed the same procedures,
was conducted after the teaching trials to select new items for generalization probes.
Baseline phase. The participant was presented with various items that were
identified from the preference assessment. The participant was asked “do you want this”
one time while the item was held in the researcher’s hand, outstretched towards the
participant. No prompting towards the correct response was provided and the item
continued to be presented for 5-sec regardless of the response given. The participant
could choose to take it out of the researcher’s hand or to not interact with the item. The
percentage of correct responding was calculated after 10 trials (i.e. one session).
Intervention. A four-step procedure consisting of errorless teaching, full-verbal
prompting, partial-verbal prompting, and no prompting was used to teach the participants
to appropriately respond to the presented stimuli.
Errorless teaching phase. The participant was presented with 10 food items in
random order. The correct response to half of the items was “yes” and to the other half
was “no.” Criteria to continue to the next phase was set as 90% or more correct
responding across two consecutive sessions.
12
“Yes” responding. These items were identified as preferred. Following the verbal
stimulus “do you want this” and the presentation of the item, the researcher immediately
said “yes” and gestured to “yes” on the visual card. Trials were counted as correct if the
participant echoed “yes” within 5-sec of the verbal and visual prompt. The correct
response was immediately reinforced with access to the item and verbal praise. Trials
were counted as incorrect if the participant engaged in a verbal response other than “yes”
or did not respond within 5-sec.
“No” responding. These items were identified as least preferred. Following the
verbal stimulus “do you want this?” and presentation of the item, the researcher
immediately said “no” and gestured to “no” on the visual card. Trials were counted as
correct if the participant echoed “no” within 5-sec of the verbal and visual prompt. The
correct response was immediately reinforced with the item’s removal and verbal praise.
Trials were counted as incorrect if the participant engaged in a verbal response other than
“no” or did not respond within 5-sec.
Prompt fading phase. The participant was presented with 10 food items in
random order. Criteria for continuation was set as 80% or more correct responding across
two consecutive sessions before the prompt could be faded to the next phase (i.e., full
echoic prompt then faded to partial echoic prompt then faded to no prompt).
Full echoic prompt. Immediately following the verbal stimulus “do you want
this?” and the presentation of the item, the researcher said “yes” or “no” depending on the
correct response. The visual card was not within sight and no gesture was made.
Following an incorrect response, the item and verbal stimulus were re-presented along
13
with the visual card which the researcher gestured to. If the participant did not respond
correctly after the trial had been re-presented, the trial was counted as incorrect and the
next trial began. Reinforcement for correct responding was access to or removal of the
item and verbal praise.
Partial echoic prompt. Immediately following the verbal stimulus “do you want
this?” and the presentation of the item, the researcher made the phonetic sound for the
letter “Y” or “N,” depending on the correct response. A full verbal “yes” or “no” was not
used, and the visual card was out of sight. Following an incorrect response, the item and
verbal stimulus were represented along with a full verbal prompt of “yes” or “no”. If the
participant did not respond correctly after the trial was re-presented, the trial was counted
as incorrect and the next trial began.
No prompt. Following the verbal stimulus “do you want this?” and the
presentation of the item, the researcher waited 5-sec and provided no prompts toward
correct responding. Following an incorrect response, the item and verbal stimulus were
represented along with a partial verbal prompt. If the participant did not respond correctly
after the trial had been represented, the trial was counted as incorrect and the next trial
began.
Generalization probes. Additional data was collected after the participant
mastered the unprompted teaching trials. The participant was presented with 10 new food
items that were identified from another preference assessment to be highly preferred or
lesser preferred. The procedures were the same as in the baseline condition.
14
Research Design
This study used an ABC across multiple baselines design. This design allowed for
the replication of baseline, intervention, and generalization phases across both
participants. The data was input into an excel document and is displayed as a line graph.
Calculation of the mean per phase and visual inspection of the overall trend, level, and
variability of the data determined if the intervention caused the change in performance.
15
Results
Figure 1 depicts Jacob and Maya’s results across baseline, intervention, and
generalization trials. Jacob was tested across two baseline sessions which resulted in an
average of 25% correct responding. During the intervention, Jacob responded correctly
99% of the time across eight sessions. In the generalization probes, Jacob correctly
responded 90% of the time across three sessions. Maya was tested across five baseline
sessions and responded correctly 2% of the time. In the prompt fading phase, Maya
correctly responded 92% of the time across 10 sessions. During the generalization probes,
Maya responded correctly 78% of the time across four sessions.
Across both participants, there is a clear level change and sharp upward trend in
the data from baseline to intervention. Without an intervention, the baseline levels of
responding would have continued to remain low. This suggests that the prompt fading
intervention was responsible for the change in Jacob and Maya’s performance. Jacob’s
data during the intervention is remarkably stable, showing little variability. Maya’s data
for the same phase is more variable, but overall stable with a slight upward trend. Jacob’s
performance during the generalization probes shows a mild level change and a slight
downward trend from intervention to generalization. Maya’s data during the
generalization probes is more drastic and displays a slight level change and steeper
downward trend across prompt fading and generalization. The data suggest that the skills
taught during the intervention maintained once prompts were removed and participants
were asked to respond to untaught items.
16
Figure 1. Jacob and Maya’s results across baseline, prompt fading, and generalization
phases.
0
10
20
30
40
50
60
70
80
90
100
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19
% C
orr
ect
Res
po
nd
ing
Sessions
0
10
20
30
40
50
60
70
80
90
100
Yes-No Manding
Jacob
Maya
Baseline Prompt Fading Generalization
17
Discussion
This study evaluated the intervention techniques of mand training and prompt
fading in discrete trial format in order to increase correct yes-no responding. Individuals
who cannot correctly respond to yes-no questions rely on others to make choices for
them. The goal of this study was to help the participants develop fundamental language
skills by teaching them to vocally respond to their environment with a clear “yes” or
“no.” Ideally, these skills would allow them to express their desires and reject what they
do not want. Overall, both Maya and Jacob’s data indicate that the intervention was
successful at increasing their yes-no manding abilities and they were able to use their new
skills when tested with untaught items.
Maya was a woman in her early 20s with limited language capabilities. She could
communicate across multiple modalities, either vocally, by manual sign, or with the
assistance of a communication binder or device. Her language mainly consisted of one-
or two-word phrases or requests. For this reason, she had received behavior and speech
therapy since she was a child. At one point, her language training included the use of a
visual yes-no card to prompt vocal responding when needed. The visual card was split in
half, one half was green with “yes” printed on it while the other half was red with “no”
printed on it. Caregivers and teachers noted that Maya was quick to respond when a
visual cue like this was present to assist her in communicating.
For this project, the researcher decided to incorporate this familiar yes-no visual
as part of the prompt fading intervention. The visual was introduced to Jacob for the first
18
time during this study’s procedures. Jacob’s responding during the intervention scarcely
varied from 100% and he required the minimum number of sessions per prompt phase.
For the errorless phase, he correctly responded to the full vocal prompt and the visual
with 100% over two sessions. In contrast, Maya needed three sessions with the full vocal
and visual prompt, and her average correct responding was 90%. Maya also needed three
sessions with the full verbal prompt and her average correct responding was 87%.
Differences in Maya and Jacob’s correct responding scores across prompt phases
could be due to a range of variables. Maya was 17 years older than Jacob and it is
possible that she was less responsive to the combination of vocal and visual prompts due
to her extended history of being prompted to communicated in ways that may have been
different than the procedures used in this study. Jacob was just starting to learn vocal
language when this project began. It is possible that he was receptive to these techniques
due to his limited exposure to other language training programs. Maya may have
struggled to overcome past prompting routines that resulted in reinforcement while Jacob
had little to no experience with yes-no mand training and reinforcement contingencies
had not been established yet.
Settings for the procedures were different across both participants. Maya’s
sessions took place in the agency’s office. She sat on a large couch in a room with the
researcher while a clinician was present to take IOA data. Jacob’s sessions took place in
his home, occasionally with a clinician present to take IOA data. His Mom was always
present in the room or within ear shot and he had access to his preferred toys while sitting
on the carpet with the researcher. Environmental stimuli varied across both of these
19
settings and without further investigation to isolate these variables, it is unclear how they
impacted performance.
The timing of sessions relative to one another may have influenced the outcome
of this study. All baseline sessions took place in one day. For Maya, intervention sessions
six through eleven took place in one day, three weeks after baseline sessions. The rest of
the intervention sessions took place the following week. Generalization sessions took
place in one day, two weeks after intervention sessions. For Jacob, intervention sessions
three through six took place in one day, one week after baseline sessions. Intervention
sessions seven and eight occurred the following week, then sessions nine and ten the
week after. Due to scheduling constraints, illness, and holidays, generalization sessions
did not occur until five weeks after intervention had concluded. The amount of time
between training and testing may have contributed to Jacobs drop in performance from an
average of 99% to 89% correct responding.
The items used in conjunction with the verbal stimulus to evoke yes-no
responding were selected based on the results of two paired stimulus preference
assessments per participant. The assessments required testing of 10 items, five were
hypothesized as highly preferred and the remaining five were hypothesized as lesser
preferred. Maya’s caregiver reported in the initial interview that she preferred food over
any other tangible item, and it was difficult to identify foods that she did not like. For this
reason, the researcher selected foods with intense flavors or textures, such as small pieces
of raw garlic cloves, ginger root, white and green onions, or raw mushrooms.
Individually tasted, Maya described these foods as “hot!” or she grimaced and drank
20
water afterwards. However, when these items were presented in quick succession during
the intervention and generalization trials, Maya appeared to hold multiple food items in
her mouth and chew them together while vocally describing them and smiling, possibly
indicating a preference for the foods once combined. This unforeseen reinforcement
contingency may have contributed to Maya’s incorrect responding. For example,
preference assessment results indicated raw mushrooms and onions as lesser preferred,
therefore “no” was selected as the correct answer. However, when asked “do you want
this?”, Maya consistently responded “yes.” Maya’s answer was speculated to have been
reflective of her desire to eat the combined food items and therefore functionally correct,
but the data did not capture this, and her responses were recorded as incorrect.
Jacob’s Mom described him as a picky eater with clear preferences for familiar
foods. The first preference assessment which selected items for the intervention trials
yielded results that appeared to accurately reflect Jacob’s preferences. The second
preference assessment results which selected items for the generalization probes was less
clear. During generalization sessions, Jacob appeared to visually inspect an animal
cracker that had been identified as a highly preferred “yes” item, before pausing and
saying “no,” annunciated in a manner different than his other yes-no responses. This “no”
was not a short and concise answer but an elongated response with vocal emphasis on the
“o.” His Mom reported that he had never eaten an animal cracker before participating in
this study. His lack of familiarity with the animal cracker and the five-week gap in
between the preference assessment, when he had last been exposed to the animal cracker,
21
and the follow up generalization sessions may have contributed to his incorrect
responding.
This study extends previous research in the field of Applied Behavior Analysis.
Research by Neef, et al. (1984) and Hung (1980) evaluated similar training techniques
but did not test if yes-no manding generalized to untaught items. This study found that
Jacob and Maya could continue to correctly mand when presented with new, untrained
items. Shillingsburg, et al. (2009) assessed a vocal prompt fading intervention across
verbal operants and evaluated participants for generalized mand responses to untrained
items. This study found the same results, confirming previous findings and validating
vocal prompt fading techniques to teach yes-no manding. This study used a visual prompt
along with a vocal prompt during the errorless teaching phase in the intervention.
Previous research has not evaluated these types of prompts together to teach yes-no
manding. The results of this study suggest that systematically faded visual and vocal
prompts across discrete trials can help teach individuals to correctly mand “yes” and
“no.”
22
References
American Psychiatric Association. (2013). Diagnostic and statistical manual of mental
disorders (5th ed.). Arlington, VA: Author.
Buron, K. D., Wolfberg, P., & Gray, C. (2014). Learners on the autism spectrum:
preparing highly qualified educators and related practitioners. Shawnee Mission,
KS: AAPC Publishing.
Cooper J. O., Heron T. E., & Heward W. L. (2007). Applied behavior analysis (2nd
ed.) Upper Saddle River, NJ: Pearson.
Hung, D. W. (1980). Training and generalization of yes and no as mands in two autistic
children. Journal of Autism and Developmental Disorders, 10(2), 139-152. doi:
10.1007/BF02408465
Jennett, H. K., Harris, S. L., & Delmolino, L. (2008). Discrete trial instruction vs. mand
training for teaching children with autism to make requests. The Analysis of
Verbal Behavior, 24(1), 69-85. doi: 10.1007/BF03393058
Kang, S., O’Reilly, M., Lancioni, G., Falcomata, T. S., Sigafoos, J., & Xu, Z. (2013).
Comparison of the predictive validity and consistency among preference
assessment procedures: A review of the literature. Research in Developmental
Disabilities, 34(4), 1125-1133. doi: 10.1016/j.ridd.2012.12.021
Kazdin, A. E. (2011). Single-case research designs: Methods for clinical and applied
settings. New York, NY: Oxford University Press.
Mangum, A., Fredrick, L., Pabico, R., & Roane, H. (2012). The role of context in the
23
evaluation of reinforcer efficacy: Implications for the preference assessment
outcomes. Research in Autism Spectrum Disorders, 6(1), 158-167. doi:
10.1016/j.rasd.2011.04.001
Michael, J. (1988). Establishing operations and the mand. The Analysis of Verbal
Behavior, 6(1), 3-9. doi: 10.1007/BF03392824
Neef, N. A., Walters, J., & Egel, A. L. (1984). Establishing generative yes/no responses
in developmentally disabled children. Journal of Applied Behavior
Analysis, 17(4), 453-460. doi: 10.1901/jaba.1984.17-453
Plavnick, J. B., & Ferreri, S. J. (2012). Collateral effects of mand training for children
with autism. Research in Autism Spectrum Disorders, 6(4), 1366-1376. doi:
10.1016/j.rasd.2012.05.008
Shillingsburg, M. A., Kelley, M. E., Roane, H. S., Kisamore, A., & Brown, M. R. (2009).
Evaluation and training of yes‐no responding across verbal operants. Journal of
Applied Behavior Analysis, 42(2), 209-223. doi: 10.1901/jaba.2009.42-209
Skinner B.F. (1957). Verbal behavior. Englewood Cliffs, NJ: Prentice
Hall, Inc.
Smith, T. (1999). Outcome of early intervention for children with autism. Clinical
Psychology: Science and Practice, 6(1), 33. doi: 10.1093/clipsy.6.1.33
Sundberg, M. L., & Michael, J. (2001). The benefits of Skinner’s analysis of verbal
behavior for children with autism. Behavior Modification, 25(5), 698-724. doi:
10.1177/0145445501255003